strava-mcp-server by iamnotreddy - MCP Server

Strava MCP Server

A project that integrates with Strava's API to provide natural language insights about your activities. I wanted to better understand the MCP (Model Context Protocol) hype by building something tangible and I've always wanted to better understand my running performance / trends in nuanced ways.

Generally I've been pretty impressed. There's nothing revolutionary about writing logic that finds your fastest activities, your longest activities, what days you typically run etc. The magic is in how MCP lets you provide AI with composable building blocks, and then watch it intelligently combine them to answer complex questions in natural language. Instead of having to dig through data visualizations or remember specific API endpoints, you can just ask "How has my training consistency changed since my injury break?" and the AI interprets your intent, orchestrates the right combination of data functions, and translates the results back into insights that actually make sense.

That said, it's not a silver bullet — you need to be intentional about designing your building blocks and their boundaries to prevent issues like API rate limits or the AI attempting to process unwieldy amounts of data. But even with those considerations, there's definitely something compelling here.

Needs some work to productionize but here's a local demo of the MCP server integrated with a geospatial interface I cobbled together. The query takes a bit to process so skip to ~1:10 to see the AI-generated insight.

https://www.loom.com/share/4d0d713a39ef4949b01710e789c088d0?sid=a2c18ea7-944c-48e8-ae05-727311624d30

Project Structure

This is a monorepo containing:

mcp-server: The core MCP server implementation + a terminal-based MCP client
api-gateway: Express service for handling API requests (not yet implemented)
shared-types: Shared TypeScript types between packages

MCP Capabilities

The MCP server provides natural language analysis of your Strava activities through several key features:

Activity Analysis

Find your fastest and longest activities
Analyze running patterns by time of day and day of week
Track monthly load progression and training consistency
Identify activity gaps and their impact on performance
Analyze double-day training patterns

Performance Insights

Calculate pace and distance trends
Compare weekday vs weekend running patterns
Analyze training consistency and progression
Track performance changes after breaks
Evaluate double-day training effectiveness

Example Questions You Can Ask

"What were my fastest 5-mile runs last year?"
"When do I typically run during the week?"
"How has my training load changed over the past 6 months?"
"What's my average pace on weekend runs vs weekday runs?"
"Have I been consistent with my training this month?"
"What's the longest gap between my runs in the last year?"
"How do my double-day runs compare to my regular runs?"

Prerequisites

Node.js (v18 or higher)
npm
Strava API credentials
Anthropic API key (for natural language processing)

Getting Started

Clone the repository:

git clone https://github.com/yourusername/strava-mcp-server.git
cd strava-mcp-server

Install dependencies:

npm install

Set up environment variables:

cp .env.example .env

Edit .env with your API keys and configuration.

Start the MCP server and client:

# In one terminal, start the MCP server
cd packages/mcp-server
npm run dev

# In another terminal, start the MCP client
cd packages/mcp-server
npm run client:dev

The MCP client provides a simple CLI interface to test the server's functionality. The Express backend is not yet implemented.

Environment Variables

See .env.example for required environment variables and their descriptions.

Development Status

This project is WIP. The core MCP server implementation is complete, while client integration and additional features are being worked on. The Express backend is planned but not yet implemented. The end goal is to connect this to Strava visualization tools I've built